In the radio communication network based on the 3rd generation partnership project (3GPP) protocol such as long term evolution (LTE), each UE initiating service request to the evolved Node B (eNB) will be set an inactivity timer which is used to track the length of the time when a UE is kept in idle state, i.e. no transmission between the UE and the eNB. Once the inactivity timer expires while the UE is still in idle state, the resources used for the UE in the eNB and the serving gateway (SGW) on the core side of the network will be released. Typically, the inactivity timer is kept by the mobility management entity (MME) and/or the eNB, and the SGW has no idea of the inactivity timer. If the inactivity timer expires, the MME or the eNB simply notifies the SGW to release the respective local resources used for the UE.
Since the eNB and the SGW release the resources used for the UE only when the inactivity timer expires, it is of great interest to set the inactivity timer as short as possible so that the occupied radio resources can be released earlier and used for other transmissions. Nonetheless, the frequent radio resources release undesirably results in a significant signaling load such as service request, handover, etc. As shown in FIG. 1, the length of inactivity timer is negatively correlated with the signaling rate. Hence, when the inactivity timer is set shorter, increasing resources in the entities of the core side such as the MME and SGW are consumed to handle the signaling. In order to make a tradeoff between the radio resources and the core resources, it is recommended to set the inactivity timer to be 61 seconds. However, more often than not, one can notice that this inactivity timer value is set to 5 seconds in practice. As such, the resources on the core side of the network are unduly occupied to handle the massive signaling resulted from the frequent radio resources release.